EVAPORATION 361 frequently that-the capacity of an evaporator is reduced to one-half or one-third by an apparently thin coating of the solids on the tube surface. For water at rest, k is about 100, but the conductance will increase rapidly with the increased velocity. Tests made by Clement and Garland (Bull. No. 40, Engineering Experiment Station, University of Illinois) show values for k of 730 at 1.45 ft. per second, and 2,500 at 17.13 ft. per second, which is considerably higher than the average conductance of boiling water. This fact is important for the construction of preheaters and surface condensers, and various experiments have given the following relations (Trans. A. S. M. E., Vol. 32): Ser, k = 520-v^v Josse, 487^7 Stanton, 340-^7 Clement and Garland, 270-^7 Hagemann, 2S2\^v; Allen, 22Q-W; Orrok, 308\/0 where v denotes the velocity of the liquid in feet per second. As stated before, the total quantity of heat transmitted depends on the tempera-1.0=t2/fii_______________________________________ .2 .3 .6 .7 .8 .9 i.0=to (for ti"=a) FIG. 1.—Heat transmission constants. tiire difference between the heating medium and the liquid. In ordinary evaporator practice, this difference is fairly constant, and can readily be determined by thermometers. In preheaters and condensers, however, the temperatures change and the average temperature difference will be ta = (t\ — t>2)/n(l — -\/U/ti) (taken from E. Hausbrand). t\ denotes the initial and t* the final temperature difference; n is a factor varying from 10 to 100 or higher, and may be taken arbitrarily, depending on the exactness of the result wanted. The results of this equation have been plotted in Fig. 1. For instance, for ti = 20° and tz = 12°, tt/ti = 0.60, and ta (for h = 1) = 0.80, therefore ta = 16°. The total sum of conductance K0 is commonly called the coefficient of heat transmission, and for steam at 212°F. and a copper tube Me in. thick, has been figured as 655. This amount will vary with the density of the steam, and empirical formulas have been proposed by various authors. Professor Kerr has determined K0 by a large